GLOSSARY
Complex
Two or more molecules bound together.
Interactor or component
One of the constituents of a complex.
Ligand
A drug-like small molecule, or the smaller of two interacting proteins.
Receptor
The larger interactor.
Binding sites
Place on the molecular surface of each interactor where it contacts the other(s).
Interface
A part of the complex where two interactors contact each other.
Docking
Identifying the structure of a complex by using knowledge of the interactors.
Rigid-body docking
A simple model which only permits wholesale translations and rotations of each interactor relative to
the other. Distances between all pairs of atoms within each interactor are not varied.
Flexible docking
A model which accommodates change in the torsion angles of the interactors. More rarely, bond
angles or bond lengths may also be varied.
Sidechain flexible docking
A model which accommodates change in the torsion angles, bond angles or bond lengths of the
sidechains of protein interactors, but not their backbones.
Protein backbone
The chain of atoms in a protein formed by the peptide bonds and the alpha-carbon atoms between
neighbouring residues.
Protein sidechains
The branches of atoms in a protein which have an alpha-carbon atom at their base.
Bond torsion
Rotating atoms downstream of a chemical bond about an axis defined by the bond.
Torsion angle or dihedral angle
The angle of bond torsion necessary to line up two reference atoms.
Bond flexing
Changing the angle between two chemical bonds formed by a given atom.
Bond angle
The angle between two chemical bonds formed by a given atom.
Native
The structure of a molecular complex as it exists in vivo. More loosely can refer to an in vitro
experimental determination by X-ray crystallography, nuclear magnetic resonance, electron
microscopy etc .
Configurations
A set of models for a molecular complex.
Decoys
A set of incorrect models for a molecular complex, constructed for the purpose of developing an
effective scoring protocol.
Double blind trial for docking (cf. CAPRI)
A trial where ‘modellers’ submit models of a given complex to ‘assessors’. Modellers do not have
access to knowledge of native structure; assessors do not have access to the identity of modellers.
Coordinate file
A computer file containing 3-dimensional coordinates for centres of the atoms of a molecule.
PDB file
The dominant file format for coordinates of proteins (‘Protein Data Bank’).
Scoring function or fitness function
A scheme for associating a number with each model of a complex aiming to reward models near to
the native complex structure.
Energy function
A scoring function which has a physical interpretation as energy.
Pairwise energy
An energy function which has a contribution from each possible pair of atoms in the model.
Electrostatics
A pairwise energy function encoding Coulomb’s law E es   q1 q 2 r12 r12  .
Dielectric
The function  r12  in electrostatics. The choice  r12   r12 is common.
Stereochemistry or Van der Waals
Penalizing geometric clashes between atoms, and rewarding glancing contact.
Lennard-Jones
A
B
A stereochemical pairwise energy function of the form E vdw   12  126  12
r
r1212
 12
Force Field

 .


A prescription for the values q1 q 2  12 A12 B12 for all atoms. Examples are CHARMM and AMBER.
Solvation
Rewarding the exposure of polar residues to water, and penalizing their burial at the interface.
Convolution
A combination of two mathematical functions based on superimposing one of them at every place
where the other is non-zero.
Fourier series
A mathematical technique based on a global sum or integral weighted by a sine or cosine.Useful for
evaluating scores which have the form of ‘convolutions’.
Monte Carlo
A generic term for mathematical techniques based on random numbers.In the Monte Carlo approach to
docking, new configurations are generated by applying random displacements to existing ones.
Normal modes
Patterns of perturbation of the atomic coordinates for which the molecule will behave like a simple
spring (with linear restoring force) in response. A protein’s normal modes with the lowest frequency of
oscillation can be identified, and tend to be similar to frequently observed patterns of large
conformational change on docking, for instance pincer and hinge movements .
Affinity
The likelihood that the interactors will associate to form the complex in the cellular environment.
Mutagenesis
An experimental technique to identify binding sites in protein complexes, by replacing each residue by
alanine and re-measuring the affinity.
Correlated mutations
Pairs of residues, one from each interactor, which are either both mutated or both not mutated across
a range of species. Binding sites are suggested by the presence of correlated mutations.
Molecular surface
The locus of points on the boundary of at least one atom and not on the inside of any atoms of the
molecule. May be artificially smoothed to remove internal voids and channels.
Solvent-accessible surface
The locus of the centre of a water molecule which is in glancing contact with at least one atom of the
molecule but does not penetrate any atoms.
Normal
The normal to at a point on a surface, is the direction in which the surface is facing at that point.
Marching cubes
Method for generating a simplicial complex which approximates a molecular surface or the solventaccessible surface of a molecule.
Voxel
A 3-dimensional pixel.
High-performance computing (HPC)
Dividing a single large task between multiple processors.
Experimental screening
Automated lab-testing a large variety of compounds for their suitability as drug leads.
Virtual screening
Using computers to score a large library of compounds for their suitability as drug leads.
ADME(T)
Acronym of the important criteria for suitability of a molecule as a drug: Absorption, Distribution,
Metabolism, Excretion, (Toxicity).
Lipinski’s rule of 5
Heuristic rules for identifying potentially ADME-compliant drugs: not more than 5 H-bond donors, not
more than 10 H-bond acceptors, molecular weight less than 500, log partition coefficient less than 5.
QSAR
Informatics methods for infer the suitability of a molecule as a drug (Quantitative Structure-Activity
Relationship).The next step in sophistication from Lipinski’s rule.